PANNS - A Parallelized Artificial Neural Network Simulator

被引：0

作者：

Fuerle, T ^{[1
]}

Schikuta, E ^{[1
]}

机构：

[1] Univ Vienna, Inst Appl Comp Sci & Informat Syst, A-1010 Vienna, Austria

来源：

PROGRESS IN CONNECTIONIST-BASED INFORMATION SYSTEMS, VOLS 1 AND 2 | 1998年

关键词：

neural network; parallel simulation system; workstation cluster; data parallelism; backpropagation;

D O I：

暂无

中图分类号：

TP18 [人工智能理论];

学科分类号：

081104 ; 0812 ; 0835 ; 1405 ;

摘要：

This paper describes the specification and implementation of PANNS (Parallel Artificial Neural Network Simulator), a general purpose tool for building, running and analyzing backpropagation neural network. This simulator allows arbitrary distribution of neural networks onto workstation clusters. The running system is based on the data parallel simulation approach. Furthermore we present an example for illustration and give performance data for a runtime analysis of the system.

引用

页码：712 / 716

页数：5